摘要
基于可靠度理论编制计算程序、对管桩承载力预测方法进行可靠度评估,并利用管桩承载力试验值与计算值之比进行可靠度分析、研究其竖向极限载力计算模型的不确定性.通过对Schmertmann和Nottingham法、Almeida等法、Powell等法、Eslami和Fellenius法、以及聚类中心法概率分布模型进行检验,发现对数正态分布的拟合特性较好、正态分布基本不被拒绝,而极值I型分布时则要慎用.模拟计算结果表明,中心点法只适用于可靠度的粗略计算、验算点法计算效率和精度可满足一般工程要求,而蒙特卡罗重要抽样法的效率较之蒙特卡罗直接抽样法有明显提高、可用于精度要求较高的工程;管桩竖向极限载力聚类中心法计算模型的可靠度要高于其他预测方法.
Based on the theory of reliability,we developed computational procedures to evaluate the predictionmethods of pipe pile bearing capacity,and then the reliability analysis was developed by the ratio of bearing capacitytested and calculated value of Pipe pile. By testing the probability distribution model with Schmertmann and Nottinghammethod,Almeida etc. method,Powell etc. method,Eslami and Fellenius method,as well as the method of clusteringcenter,we concluded that the logarithmic normal distribution fits well,the normal distribution is not rejected at large,however,the extreme type I distribution must be careful to apply. Our analysis and simulation results displayed that thecenter point method only is suitable for rough calculation of reliability,accuracy of checking point method meets thegeneral engineering requirements,the efficiency of Monte-Carlo important sampling method is increased significantlycompared with Monte-Carlo direct sampling method,and thus may be used for high precision engineering projects. Westudy the uncertainty of vertical critical residual calculation model by the load ratio,it suggested that the reliabilityof clustering center method is higher than other prediction methods.
出处
《河南科学》
2016年第4期542-547,共6页
Henan Science
基金
国家十二五科计支撑计划项目(2012BAJ01B04)
国家自然科学基金项目资助(51508521
51278364)
关键词
管桩
承载力
预测方法
可靠度
荷载比
聚类中心
reliability
pipe pile
bearing capacity
predication method
load ratio
clustering center
